FlexiBMS Lite - Flexible Configuration BMS w/ CAN-bus

That is number rounding behavior for the terminal printing, as it only prints out 4 decimals, so in the internal memory the actual value is likely something like 1.01309272, but the terminal hard cuts the decimals after the 4th, so that’s why it only shows 1.0130 when you put in 1.0131, which gets rounded down to something like 1.01309xxxx.

Any NTC-resistor with a 10kOhm resistance at 25C can be used, but check and configure the correct Beta-value.

That is a bit weird, I checked the store page and didn’t see the beta-value mentioned anywhere… nor a datasheet.

quick googling with the product number got a hit on a forum, which said that it has a Beta-value of 4100. Datasheet link is dead.

2 Likes

Maybe you could offer to buy NTC resistor as a extra on your website? Seems several people would have payed. Maybe a 2pin header and and plug. (Could this be incorporated in to the 2nd balance plug on the Flexi BMS 18s?)

1 Like

I’m making slight change to the 12S so it would fit a vertical JST-PH-2 (2.00 mm pitch, the pitch on the OG model is 2.50mm) for the NTC-probe neatly.

And as the board is getting longer on the 18S, I believe there is space to have independent connectors for both NTC-probes on the charger port side.

2 Likes

Hey @SimosMCmuffin, when will the flexi be available again? The OG 12s version i mean. :slight_smile:

@Linesflag STM32 Chip is on back order for a while

1 Like

@SimosMCmuffin which parts are missing? Maybe we can crowdcheck if we find it somewhere?

1 Like

I connected my Flexi to a new 10S2P p42a battery pack for the first time.

Now I’m noticing a noise from the bms.
With my iphone I registered a track of ~38dBA, I don’t know if it’s useful :

It’s normal? :sweat_smile:

The OG is getting updated to the Lite 12S, with the slight component upgrades for higher specs.

The STM32L433CBU6 is currently absolutely unobtainium. It doesn’t exist in stock, anywhere.

Even in STM’s own store it’s out-of-stock.

I also checked for the bigger flash version, STM32L433CCU6 with 256KB flash (CBU6 has 128KB flash) and the higher temperature rated one, STM32L433CCU3 instead, but both are also unobtainium.


So, unfortunately, and also to my big disappointment, the next build batch is completely on hold. Sorry.

I took inventory and I have 5 units of the MCU, but I think I’ll hold on to those for any possible repairs.

I also looked into different packaging of the MCU, like the TQFP- and WLCSP-packages, but they are also very low on availability. The TQFP, would require a bigger PCB, as it needs more real estate and to be able to use the WLCSP needs BGA level HDI (High Density Integration) for the board design and that requires some special pcb fabrication requirements.

The WLCSP-package with 49 balls (7x7 grid) would be smaller at ~3.3x3.3mm outline compared to the 48-pin QFN package at 7x7mm outline, but the pitch is an insane 0.4mm (distance between the balls on the bottom) that it requires a specialized PCB design and I have attached a video below that demonstrates the insanity of such a small package, with such many balls.

That is an inductor whine from the buck-regulator and it’s normal for this application with the used components.
The inductor’s magnet wire is slightly vibrating with the ON-OFF pulses (EMF causes a physical force to be enacted on the wire) and it has just enough movement that it transmits into the surrounding air, which in this case is a high pitch whine. During the boot-up sequence with the color sequence, you can hear the pitch of the whine slightly change with each different color on the LED, as the current flow changes ever so slightly between the different color diodes being turned on.

6 Likes

Yes, I was aware of that. I asked because of the pitch that seemed to be pretty loud.

Now that I’ve updated the firmware to v0.18 the noise is significantly lower (almost inaudible), how is it possible?

I’ve also another question: if I wanted to wire the CAN cable without the opto-enabler (only CANH and CANL), what should I change beyond the parameter $40=1 ?
I’ve read the wiki and If I’m not wrong it seems to be not updated for the latest release.

I’ll attach my parameters configuration if useful.

parameters.rtf (4.7 KB)

Here are the settings for the Davega X maybe this helps FlexiBMS Lite Setup - DAVEGA

3 Likes

This is your solution @edoardoconti

1 Like

Hey @SimosMCmuffin, I’m getting some weird numbers from status reporting here. The first line shows how the cell voltage actually is (confirmed by measuring with a multimeter), then they start to fluctuate wildly. Funnily enough, the combined voltage stays correct, only cell level is weird. Also the temperature readings are jumping like crazy. I’m running 0.18 FW.

Update: Now its not doing the first line thing anymore, instead the correct voltages show up in a line somewhere into 5-10 seconds of reporting.

Flexi Is working! Never seems to get the pack balance trimmed all the way though, what can I do to fix this?

2 Likes

0.01v mismatch is totally ok :slight_smile:

1 Like

Yeah, same with LLT smart bms.

To which point do you get it? Best I was able to do was about 0.005v on the LTT.

Best I could get with one pack was precisely 0.01V.

The standard settings are 0.01V you can manually change it to be more precise but it gets super ineffective after that so never mind. just so you know. It works with the 3rd download on their website on android for me.

2 Likes

I tried to put less than that but I think it didn’t accept the value as when I went again to the conf screen it was 0.01 again :man_shrugging:

You need to write the new settings with the checkmark in the top right corner.